RELATE: Physically Plausible Multi-Object Scene Synthesis Using Structured Latent Spaces

TL;DR

RELATE is a novel generative model that produces physically plausible multi-object scenes and videos by explicitly modeling object correlations, enabling realistic scene editing and outperforming prior methods on synthetic and real-world data.

Contribution

It introduces a physically interpretable, object-centric generative model that captures object correlations for improved scene and video synthesis.

Findings

Outperforms prior art in object-centric scene generation

Enables realistic scene editing

Extends naturally to dynamic scenes and videos

Abstract

We present RELATE, a model that learns to generate physically plausible scenes and videos of multiple interacting objects. Similar to other generative approaches, RELATE is trained end-to-end on raw, unlabeled data. RELATE combines an object-centric GAN formulation with a model that explicitly accounts for correlations between individual objects. This allows the model to generate realistic scenes and videos from a physically-interpretable parameterization. Furthermore, we show that modeling the object correlation is necessary to learn to disentangle object positions and identity. We find that RELATE is also amenable to physically realistic scene editing and that it significantly outperforms prior art in object-centric scene generation in both synthetic (CLEVR, ShapeStacks) and real-world data (cars). In addition, in contrast to state-of-the-art methods in object-centric generative modeling, RELATE also extends naturally to dynamic scenes and generates videos of high visual fidelity. Source code, datasets and more results are available at http://geometry.cs.ucl.ac.uk/projects/2020/relate/.